Skip to content

About

Veklom BYOS Backend — Private governed AI backend. Run AI workloads with policy, routing, cost controls, audit/evidence, API keys, tenant isolation, and compliance on your own server or cloud.

Topics

Resources

Stars

1 star

Watchers

0 watching

Forks

Latest commit

 

History

1,402 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Important

VEKLOM BIBLE — READ FIRST: 00_VEKLOM_BIBLE.md It supersedes older cross-repo architecture, deployment, topology, port, and agent-alignment claims wherever they conflict.

Veklom BYOS Backend

Veklom AI is a private governed AI backend that allows secure, compliant, and customizable AI workload management on your own infrastructure, offering policy control, cost management, auditing, and tenant isolation.

Core Capabilities

  • Private AI Backend: Veklom BYOS (Bring Your Own Server) allows organizations to run AI workloads on their own servers or cloud infrastructure, ensuring full control over data and deployment environments.
  • Governance and Policy Control: The platform provides detailed policy management capabilities to enforce organizational standards and compliance requirements. Teams can define rules on how AI models are used, which workloads are permitted, and what access controls are applied.
  • Routing and Multi-Tenant Support: Veklom supports workload routing and tenant isolation, enabling multiple teams or clients to operate securely on the same infrastructure without risk of data crossover.
  • Cost Controls: The platform includes cost management features that allow monitoring and controlling compute expenditure for AI workloads across tenants, helping organizations optimize resource usage and reduce unnecessary spend.
  • Audit and Evidence Tracking: Every AI operation can be logged for audit purposes. This ensures accountability, traceability, and evidence collection for regulatory compliance or internal reviews.
  • API Key Management: Veklom provides flexible API key management, making it straightforward to integrate with existing applications and control access to AI models and services.
  • Compliance and Security: The platform emphasizes regulatory compliance, supporting industry-standard security practices and allowing organizations to maintain audits, access controls, and secure operations on their own infrastructure.

Deployment Flexibility

Veklom can be deployed on local servers or in private clouds, providing enterprises with maximum control over data location, security, and network configuration. This flexibility is especially valuable for organizations dealing with sensitive data or operating in highly regulated industries such as healthcare and finance.

Summary

Veklom is designed for organizations that need a secure, multitenant, and policy-driven AI platform while maintaining full control over their infrastructure. Its key features—policy management, tenant isolation, cost controls, audit logging, and API integration—make it suitable for enterprise-grade AI operations where governance, compliance, and operational transparency are critical.


Quick Start

git clone https://github.com/reprewindai-dev/veklom-byos-backend
cd veklom-byos-backend
cp .env.example .env
# Fill in required env vars (see ENVIRONMENT.md)
pip install -r requirements.txt
alembic upgrade head
uvicorn backend.apps.api.main:app --reload

Health check: GET http://localhost:8088/health

Login & Registration

The backend supports three authentication methods:

  1. Register - Create a new account:

    POST /api/v1/auth/register
    {
      "email": "your@email.com",
      "password": "your-password",
      "full_name": "Your Name",
      "workspace_name": "My Workspace"
    }
  2. Login - Sign in with email/password:

    POST /api/v1/auth/login
    {
      "email": "your@email.com",
      "password": "your-password"
    }
  3. Free Evaluation Session - Try without registration:

    POST /api/v1/auth/eval-session
    {
      "fingerprint": "browser-fingerprint"
    }

After login, use the access_token in the Authorization: Bearer <token> header for authenticated requests.


Documentation

Doc Purpose
SETUP.md Local development setup
DEPLOYMENT.md Docker / Coolify / Render / Hetzner deploy
ENVIRONMENT.md All environment variables explained
API_SURFACE.md Every route family and endpoint
SECURITY_MODEL.md Auth, isolation, secrets architecture
LICENSE_ACTIVATION.md Buyer license key activation
BUYER_PACKAGE.md What you receive, support terms
SELLABLE_BACKEND_AUDIT.md Module readiness audit
SOURCE_BACKEND_INVENTORY.md Full folder/route/dep inventory
DETERMINISTIC_AI_INFRASTRUCTURE.md Veklom/UACP/GPC/py03-irongrid role doctrine
GOVERNED_OPERATIONAL_RUNTIME.md Operational runtime substrate, telemetry, governance, BYOS, and survivability spec

Architecture

veklom-byos-backend/
├── backend/
│   ├── apps/api/
│   │   ├── main.py          # FastAPI app, middleware, router registration
│   │   └── routers/         # 43 router modules
│   ├── core/                # Config, auth, DB engine, security
│   ├── db/                  # SQLAlchemy models, Alembic migrations
│   ├── license/             # License server, package guard
│   ├── scripts/             # Deploy, health, license scripts
│   └── tests/               # pytest test suite
├── Dockerfile
├── docker-compose.yml
├── requirements.txt
├── .env.example
└── .env.production.example

Core Capabilities

AI Execution

  • POST /api/v1/exec — SSE streaming inference with policy enforcement
  • OpenAI-compatible interface
  • Connect to OpenAI, Anthropic, vLLM, Ollama, or any compatible endpoint
  • Per-request cost prediction before execution
  • Autonomous routing: selects best model by cost/quality/risk

Governance & Compliance

  • Content safety scoring on every request
  • PII/PHI detection and optional redaction
  • Compliance regulation checks (HIPAA, GDPR, SOC2)
  • Explainability endpoint for model decisions
  • Tamper-evident audit logs with SHA hash chains

Security

  • JWT authentication + optional MFA
  • API key management with scoped permissions
  • Kill switch: instantly revoke all AI access
  • Locker isolation: per-tenant security boundaries
  • Security event log

Billing & Cost

  • Token wallet with prepaid credit model
  • Budget rules with hard/soft limits
  • Real-time spend tracking
  • Stripe subscriptions + metered usage
  • Topup checkout flow

Multi-Tenant

  • Full workspace isolation
  • Role-based access (owner, admin, member, viewer)
  • Per-workspace model configs, API keys, budgets
  • Admin panel for user/workspace management

License

Commercial license. See LICENSE_ACTIVATION.md and BUYER_PACKAGE.md.

© 2026 CO2 Router / Veklom. All rights reserved.

About

Veklom BYOS Backend — Private governed AI backend. Run AI workloads with policy, routing, cost controls, audit/evidence, API keys, tenant isolation, and compliance on your own server or cloud.

Topics

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Used by

Contributors

Languages